Alert button
Picture for Benjamin Van Roy

Benjamin Van Roy

Alert button

Stanford University Department of Electrical Engineering

Fine-Tuning Language Models via Epistemic Neural Networks

Nov 03, 2022
Ian Osband, Seyed Mohammad Asghari, Benjamin Van Roy, Nat McAleese, John Aslanides, Geoffrey Irving

Figure 1 for Fine-Tuning Language Models via Epistemic Neural Networks
Figure 2 for Fine-Tuning Language Models via Epistemic Neural Networks
Figure 3 for Fine-Tuning Language Models via Epistemic Neural Networks
Figure 4 for Fine-Tuning Language Models via Epistemic Neural Networks
Viaarxiv icon

On Rate-Distortion Theory in Capacity-Limited Cognition & Reinforcement Learning

Oct 30, 2022
Dilip Arumugam, Mark K. Ho, Noah D. Goodman, Benjamin Van Roy

Viaarxiv icon

Is Stochastic Gradient Descent Near Optimal?

Oct 06, 2022
Yifan Zhu, Hong Jun Jeon, Benjamin Van Roy

Figure 1 for Is Stochastic Gradient Descent Near Optimal?
Figure 2 for Is Stochastic Gradient Descent Near Optimal?
Figure 3 for Is Stochastic Gradient Descent Near Optimal?
Figure 4 for Is Stochastic Gradient Descent Near Optimal?
Viaarxiv icon

Robustness of Epinets against Distributional Shifts

Jul 01, 2022
Xiuyuan Lu, Ian Osband, Seyed Mohammad Asghari, Sven Gowal, Vikranth Dwaracherla, Zheng Wen, Benjamin Van Roy

Figure 1 for Robustness of Epinets against Distributional Shifts
Figure 2 for Robustness of Epinets against Distributional Shifts
Figure 3 for Robustness of Epinets against Distributional Shifts
Figure 4 for Robustness of Epinets against Distributional Shifts
Viaarxiv icon

Ensembles for Uncertainty Estimation: Benefits of Prior Functions and Bootstrapping

Jun 08, 2022
Vikranth Dwaracherla, Zheng Wen, Ian Osband, Xiuyuan Lu, Seyed Mohammad Asghari, Benjamin Van Roy

Figure 1 for Ensembles for Uncertainty Estimation: Benefits of Prior Functions and Bootstrapping
Figure 2 for Ensembles for Uncertainty Estimation: Benefits of Prior Functions and Bootstrapping
Figure 3 for Ensembles for Uncertainty Estimation: Benefits of Prior Functions and Bootstrapping
Figure 4 for Ensembles for Uncertainty Estimation: Benefits of Prior Functions and Bootstrapping
Viaarxiv icon

Deciding What to Model: Value-Equivalent Sampling for Reinforcement Learning

Jun 04, 2022
Dilip Arumugam, Benjamin Van Roy

Viaarxiv icon

Between Rate-Distortion Theory & Value Equivalence in Model-Based Reinforcement Learning

Jun 04, 2022
Dilip Arumugam, Benjamin Van Roy

Viaarxiv icon

Nonstationary Bandit Learning via Predictive Sampling

May 04, 2022
Yueyang Liu, Benjamin Van Roy, Kuang Xu

Figure 1 for Nonstationary Bandit Learning via Predictive Sampling
Figure 2 for Nonstationary Bandit Learning via Predictive Sampling
Viaarxiv icon

Sample Complexity versus Depth: An Information Theoretic Analysis

Apr 07, 2022
Hong Jun Jeon, Benjamin Van Roy

Figure 1 for Sample Complexity versus Depth: An Information Theoretic Analysis
Figure 2 for Sample Complexity versus Depth: An Information Theoretic Analysis
Viaarxiv icon